文档介绍：第 31卷第 12期仪器仪表学报 V o l1 31 N o1 12
2010年 12月 Ch inese Journa l o f Sc ientific Instrum ent D ec. 2010
基于图像处理的三峡库区水域水华检测
算法及预警研究*
石为人, 贾承晖, 王楷
(重庆大学自动化学院重庆 400044)
摘要: 针对传统水华监测方法存在的过分依赖人工操作、实时性差、成本高、普及性差、主观性大、劳动强度大等缺陷, 提出了
一种可以自动检测三峡库区水域水华现象的图像处理算法, 并将该算法应用于库区水域水华预警。即利用历史数据建立颜色
先验模型, 运用改进的 Cam sh ift算法检测水华区域, 通过区域面积的变化分析水华灾情趋势, 并建立水华预警模型进而对水华
预警。实验结果表明该算法应用于水面视频监控可以自动检测出水华区域并有效地对水华实时预警。
关键词: 水华检测; 图像处理; 改进的 Cam sh ift算法; 水华预警; 三峡库区
中图分类号: TP391. 41 文献标识码: A 国家标准学科分类代码: 510. 4050
Research on water bloom detecting algorithm and early warning based
on im age processing for Three es R eservoir area
Sh iW eiren, Jia Chenghu,i W ang Kai
(C ollege of A utom ation, Chongq ing Un iversity, Chongqing 400044, China)
Abstract: Trad itionalw ater bloom detecting m ethods have the features such as over-re liance on m anual operation,
low rea-l tim e, high cos,t poor popu larity, large subjective, labor-intensive and o ther de fects. A im ing at these de-
fects, an algor ithm based on im age processing technology was proposed, w h ich can autom atica lly detectw ater b loom s
in Three Gorges R eservo ir A rea, and the a lgorithm w as app lied to w ater bloom early w arn ing in the reservo ir area.
The algorithm uses h istorical data to estab lish a co lor prio rm ode,l usesm odified Cam shift a lgorithm to detect the re-
gion o fw ater bloom s, obta ins disa